Do you know readers of your papers ?
Or - is it true that if paper is not groundbreaking no one (except very few people, working very closely and hence known for you) will read it ?
If it is so - you can write "from expert to experts" concentrating on details. If it not so - it is better to include more introduction, background, motivation, examples.
"Definition" READER is any person looking at yours paper will extract something useful for him. That is why it is difficult to measure by citations number - many papers are cited but not "readed" and vice versa.
November 7, 2011 at 3:21 pm ... I once heard that the average number of readers of a mathematics paper is less than 1."
I heard the same from other people - this is based on impact factor which they say averagely less than 1 for most of math journals.
To my taste "ideal" paper should be written in such a way that just "undergrad" can understand it. But that would of course require enormous efforts and size. Which is unjustified if you know that no one will read it :)
Practical situation - I am now writing a paper, which I am almost sure some 5 experts (whom I know) will look. The whole number of experts worked quite closely around the topic in last 10 years is may be bigger - 10-20 (depends on word "closely"). But different people has different tastes and interests, so I am not sure all of them will look, moreover I feel from previous experience many will not.
On the other hand, I feel, that some questions discussed in the paper are at least understandable for any undergraduate (and hopefully interesting). So potential number of readers is much bigger - that is why I am spending time to write it as readable as I can do. But it takes much time. It does not seems to me worth to spend so much time if I know that only those 5 experts (whom I know) will look on it.
Should I spend this time - or just put formulations of the theorems and write "obviously" ? :) (For experts this would be Okay :)